Why So Many Clever Pilot Programs Flunk Out at Scale
“Small experiments for solving social problems may seem to work, but at least half of them fall apart when they’re expanded to a larger constituency. Costs are the main explanation, although not the only one.”